[cups.general] Re: SNMP printer discovery documentation?

[Kurt Pfeifle]

wtautz <wtautz at cs.uwaterloo.ca> wrote (Thursday 27 July 2006 18:48):

> Kurt Pfeifle wrote:


>>     http://www.kdedevelopers.org/node/2126

>>     http://www.kdedevelopers.org/node/2127

>>     http://www.kdedevelopers.org/node/2128

>>     http://www.kdedevelopers.org/node/2129
>>   
> Wow, more Ubuntu twists. I hadn't realized that one actually
> has to put in the snmp backend soft link.

Well, it really is a distro tweak (may be the same in Debian proper).

> It works! 

Yes. And I think it was really stu^H^Hhortsighted to disable it.

The reason given was that "it was too late introduced" and so
Dapper could not ship it because it was a new feature. 

Looks like using an arbitrary Subversion checkout (Dapper Alphas 
and Betas included the same CUPS Subversion revision "r4929" for 
several months) at a time when there was not yet even an official 
Beta of CUPS 1.2 is fine -- but when CUPS matures & makes official
releases and officially ships some features it is also OK to just 
chop off those features for being "too new and untested"....

> I gather that snmp discover only works for the same subnet that
> the print server is on? 

Yes and no.

The default setting is to use broadcasts, and broadcasts (usually)
are not transmitted into another subnet by routers and gateways.

However, I found when I individually listed all (or only the
suspicious) IP addresses from a neighboring subnet like this 
inside snmp.conf

  Address 10.162.5.1
  Address 10.162.5.2
  [...]
  Address 10.162.5.254

the snmp backend found printers there just fine when run from the
commandline. Still need to test it via the web interface as well.

> Does this snmp discover process also 
> allow for automatic determination of whether the printer
> has a duplexer?

Not yet, but it may be possible to succeed with that in a future
version.

Cheers,
Kurt